1. Field of the Invention
This invention relates generally to technologies for switching and routing optical wavelengths. More particularly, this invention relates to innovative method, structures and processes to manufacture and design improved waveguide grating-based wavelength selective switches.
2. Description of the Related Art
Current state of the art in wavelength-selective optical switching based signal transmission systems are still limited by several performance deficiencies caused by crosstalk, low coupling efficiency, and large size and poor form factors.
Because of the extremely wide transmission bandwidth allowed by optical fibers, all-optical fiber networks are increasingly being used as backbones for global communication systems. To fully exploit the fiber bandwidth in such networks, wavelength-division multiplexing (WDM) and wavelength-division demultiplexing (WDD) technologies are employed so that several independent optical signal streams may share the same fiber simultaneously, wherein the streams are distinguished by their center wavelengths. In the past, the adding, dropping, and cross connecting of individual optical signal in communication systems are done by first converting the optical signal into electrical signals. The electrical signals are manipulated electronically, which are then converted back into optical signals. However, the development of all-optical WDM communication systems has necessitated the need for all-optical wavelength selective devices. It is desirable for such devices to exhibit the properties of low insertion loss, insensitivity to polarization, good spectral selectivity, and ease of manufacturing.
In general, there are three prevailing types of technology commonly implemented in the all-optical Dense WDM (DWDM) networks: (1) Thin Film Filter (TFF), (2) Arrayed Waveguide Gratings (AWG), and (3) Fiber Bragg Grating (FBG). Among these three types of implementations, TFF technology is the predominant choice when the channel spacing is greater than 100 GHz. The advantages of TFF-based devices are that they are relatively insensitive to temperature, have minimal cross talk, and provide good isolation between different wavelengths. However, devices built using current TFF technology have the following disadvantages: they are difficult to manufacture when the channel spacing is below 200 GHz; the packaging cost is very high; and the yield is low. Due to these disadvantages, when the channel spacing is 100 GHz or less, AWG and FBG wavelength selecting devices dominate the market. The advantages of AWG devices are they can support high channel counts, are easy to manufacture, and have a small footprint. Meanwhile, the disadvantages are that AWG devices are prone to cross talk and their packaging is complex. The second dominant technology, i.e., the FBG technology, has the advantages of short development time, low capital investment, and low packaging cost as channel spacing is reduced to 100 GHz or less. However, the FBG products available in the current market have relatively high loss. Moreover, each channel requires a circulator, which increases component costs and possibly increases packaging costs.
Furthermore, there are several optical switching technologies under development today. They are as follows: Micro Electro-Mechanical Systems (MEMS), Liquid Crystals, Thermal Optics, Holograms, Acousto-Optic, etc. Among all these optical switching technologies, MEMS is emerging to be the most promising technology, as benefited from its potential of batch processing and cheap replication, as well as its sound record on reliability in a wide range of applications. All the other technologies are still in the experimental stage and need years to become reliable enough for commercial applications. Different embodiments of MEMS optical switches are made available in the marketplace that are implemented with a de-multiplexing device to first separate the input signals into multiple channels (each having a specific central wavelength) transmitted over a specific waveguide. Optical switching operations are performed for each of these de-multiplexed signals. Then a device is employed to multiplex these switched signals for transmission over optical fibers. Alternately, the wavelength selective optical switches are implemented with a de-multiplexing device to first separate the optical signal into channels of different wavelengths. The optical switching operations are carried out for each channel and these channels are connected to optical output ports. Again, a de-multiplexing operation must be performed first before wavelength selective switching can be carried out.
There are two types of optical MEMS switch architectures under development, or commercially available: mechanical and micro-fluidic. Mechanical-type MEMS-based switches use arrays of miniaturized mirrors fabricated on a single chip. The optical signal is reflected off this tiny mirror in order to change the transmission channel. Micro-fluidic-type MEMS-based switches, on the other hand, have no moving mirrors. Rather, they rely on the movement of bubbles in micro-machined channels.
Mechanical-type MEMS-based switches can be further classified into two categories according to mirror movement: two-dimensional (2-D) switches and three-dimensional (3-D) switches. In 2-D switches, the mirrors are only able to execute a two-position operation—that is, the mirrors can move either up and down or side by side. In 3-D switches, the mirrors can assume a variety of positions by swiveling in multiple angles and directions. These products (2-D switches or 3-D switches) are able to offer such benefits as excellent optical performance, minimal cross-talk, and the promise of improved integration, scalability, and reliability. On the other hand, these products and their methods of use are disadvantageous in the following aspects, First, in these switches, light travels through free space, which causes unbalanced power loss. Secondly, in order to steer each mirror, three to four electrodes need to be connected to each mirror, which is a major challenge to produce large-scale mechanical-type MEMS-based switches. Thirdly, alignment and packaging are difficult tasks particularly for large-scale switches.
While above-mentioned micro-mirror-based approach is widely taken by most major companies to build up their MEMS-based optical switches, Agilent Technology, Inc. has developed micro-fluidic-type MEMS-based switches by combining its micro-fluidics and ink-jet printing technology. In these switches, an index-matching fluid is used to switch wavelengths. This fluid enables transmission in a normal condition. To direct light from an input to another output, a thermal ink-jet element creates a bubble in the fluid in a trench located at the intersection between the input waveguide and the desired output waveguide, reflecting the light by total internal reflection. The advantages of these switches are that they have no moving mechanical parts and are polarization independent. The disadvantages are their questionable reliability and the excessive insertion loss for large-scale switches.
A common drawback of both of these MEMS-based switches is the requirement to work with external de-multiplexing and re-multiplexing devices in order to function properly in an optical networking system. The requirements of implementing de-multiplexing and re-multiplexing functions add tremendous complexities to the system configuration and significantly increase the cost of manufacture, system installation, and maintenance of the optical network systems. Another drawback is that these prior-art switching systems are not wavelength selective switches. In another words, switching systems based on MEMS cannot selectively switch a particular wavelength from an input waveguide to a desired output waveguide. In short, they are not wavelength intelligent devices.
To add wavelength intelligence to optical switches, Bragg grating has been shown to have excellent wavelength selection characteristics. A Bragg grating behaves as a wavelength-selective filter, reflecting a narrow band of wavelengths while transmitting all other wavelengths. The Massachusetts Institute of Technology (MIT) has developed a technology for building Bragg grating devices in planar optical waveguides. These so-called integrated Bragg gratings offer many advantages over the fiber Bragg grating, according to MIT. However, since switching optical signals requires routing of the optical transmission, a wavelength-selective filter alone is not sufficient to build an optical switch. In most waveguide grating device designs, since input signal and its reflective response occur within the same waveguide (see FIG. 1), circulators must be employed. One of the drawbacks of using circulators is that the circulators are expensive thus unduly increasing the implementation cost. For that reason, the grating assisted directional couplers (GADC) such as those shown in FIG. 2 are used. The optical performance characteristics of CADC and their dependence on material properties of the transmission medium are well-known, e.g., “Contra-directional coupling in grating-assisted guided-wave devices,” Hong et al, IEEE Journal of Lightwave Technology, pp. 873, vol. 10, no. 7, July 1992. Typically, four distinct wavelengths participate in the coupling process. Among these four wavelengths, only three of them are significant in our discussion. These 3 wavelengths are illustrated in FIGS. 3A and 3B. As shown in these drawings, the first response is reflection along the incident direction the second response is reflection along the coupling waveguide (contra-directional propagation) and the third response is the forward transmission along the coupling waveguide (co-directional propagation). Based on the utilizations of these optical responses between the grating assisted coupling waveguides, many different devices according to different combination of configurations are disclosed for varieties of application in controlling, switching and managing signal transmissions in optical networks.
Specifically, Horita et al. disclosed in U.S. Pat. No. 5,859,941, entitled “Optical Add/Drop Multiplexer Device”, an add/drop multiplexer device provided to extract or inserting optical signals or arbitrary wavelength and having a wavelength selection characteristic with a narrow bandwidth. The multiplexer device is formed on a substrate implemented with a layered structure. In another U.S. Pat. No. 6,567,573, entitled “Switchable Optical Components” issued on May 20, 2003, Domash et al. disclosed electrically switchable Bragg grating devices and device geometries realized by using holographic polymer or dispersed liquid crystal materials.
However, these patented technologies have technical limitations due to the interference of signal transmissions may occur when the wavelength of the reflected waves falls within the optical communication band as will be further explained below in FIG. 4. As a result, performance degradation of the telecommunication networks is possible if not properly managed.
Therefore, a need still exists in the art to provide an innovative configurations and method of manufacturing and designing the wavelength selective optical devices to overcome such limitations.